Kamal Jabbour

Spanning multiaccess channel hypercube computer interconnection

A distributed computer system based on a hypercube topology is proposed, where multiaccess channels spanning all dimensional axes provide processor interconnection. The multiprocessor system can be built at relatively low cost by combining the recent advances in high-capacity channels available through fiber optics, and demand-assignment multiple-access protocols, creating a highly fault-tolerant system with good average packet delay and average distance characteristics, at a significantly reduced cost. A totally distributed system has been achieved in which the packets are self-routing, able to take advantage of the multiple paths of equal length between two modes. The proposed structure gives a highly fault tolerant system, degrading gracefully in the presence of failures. >

A Unified Approach to Local Area Network Interconnection

An approach to local area network interconnection is presented Which combines the advances in static interconnection topologies, demand assignment multiple access protocols, and the availability of high-bandwidth fiber optic channels to create a cost-effective structure capable of interconnecting a large number of LANs with heavy traffic. This approach is independent of the protocol implemented at each LAN. The structure is based on a hypercube topology where each vertex of the graph represents a LAN. Multiple access channels spanning all dimensional axes are used in this scheme. This approach is compared to a topology with direct point-to-point connections between all nodes sharing a common axis. Through the development of the degree, diameter, average distance, cost, and average packet delay, we show that using fewer high-capacity channels, a LAN interconnection network with excellent performance characteristics can be constructed, able to support a large number of LANs with heavy traffic at a significant reduction in cost over the point-to-point case. The resulting structure has many of the desirable characteristics for static interconnection networks such as high fault tolerance, totally distributed packet routing in the interconnection network, low average distance for good performance, and low degree, resulting in low cost. For the total number of required LAN nodes and the expected amount of internode traffic, the structure is optimized for minimum cost.

Formal specification and verification of communication protocols using automated tools

The paper compares and contrasts various methods presently available for specification, validation and verification, with emphasis on verification. We describe an application of formal methods to protocol specification, validation, and verification, using an actual protocol as an example. We use a process algebra to build models of an OSI protocol and test these models using a model checker. Model checking allows us to verify large and complex models and use formal methods as practical solutions. We also show an example with an automated theorem prover.

Formal specification and verification of OSI session layer protocol using CCS

This paper describes an application of formal methods to protocol specification, validation and verification. The chosen formal method is the Calculus of Communicating Systems (CCS), which is used to build a CCS model of the Open Systems Interconnection (OSI) Reference model session layer protocol and prove its inherent livelock using the CCS automated support.<<ETX>>

Hypercube-based local lightwave networks

The authors propose a ring topology, connected by an optical fiber medium, to implement the routing functions supported by the hypercube topology. They implement the hypercube topology on a local lightwave network using wavelength-division multiplexing of optical fibers. A simple static routing scheme is used which allows for concurrent interprocess communications. The proposed network provides the necessary bandwidth to perform distributed and parallel computing over local area networks. The hypercube-based local lightwave network implemented on a multichannel ring topology is shown to require fewer channels than similar size ShuffleNet networks. The network architecture and routing scheme are presented. Possible applications are discussed.<<ETX>>

An optical hypercube local area network

We present an optical hypercube local area network which employs both free-space optics as well as optical fiber interconnections. Each node of the hypercube consists of an optical routing switch (ORS) to which users connect through an optical interface unit (OIU), providing a hierarchical structure. Self electrooptic effect devices (SEEDs) and two-dimensional spatial light modulators (SLMs) are used do perform high speed packet switching in the optical domain, providing an all-optical path between source and destination nodes. Electrooptic conversions occur only at the source and destination, eliminating the electronic bottleneck at the intermediate nodes. The network requires no wavelength division multiplexing and can be implemented using currently available optical technology to achieve high-speed switching and routing with low power requirements.<<ETX>>

An expert system approach to load balancing in a distributed environment

Distributed computing systems composed of heterogeneous and autonomous components are continuously growing in their size, complexity and diversity. Resource utilization, such as the CPU and the communication network, vary dynamically. In addition, the quality of service requirements vary from one application to the other. These characteristics necessitate adaptive and dynamic solutions for the load balancing problem. In this research, we are investigating knowledge-based techniques for developing an expert system environment, in which heuristic knowledge and real-time decision making are used to manage such complex and dynamic systems in a decentralized fashion.<<ETX>>

An optical network interface unit for multichannel ring networks

The authors present an all-optical hypercube-based multichannel network, implemented on a ring topology, using wavelength division multiplexed channels. For a network of size N, only log/sub 2/N distinct channels are required. They design an optical network interface unit (ONIU), which employs free-space optics to perform routing and switching functions. The design eliminates the electronic bottleneck at intermediate nodes and is capable of performing high speed packet switching. ONIU can be reconfigured in the GHz range and provides the high bandwidth necessary to perform distributed and parallel computing in a local area network environment.<<ETX>>

An expert system for network management

The authors propose a generalized architecture for an expert system to manage the resources of a computer network and/or a distributed system. The authors apply the proposed architecture to construct an expert system for allocating the resources needed by a task in a distributed computing environment. The resources are allocated such that the system load and the average delay of executing a task are optimized. The authors present an approach to evaluate the average delay incurred during the execution of a distributed task that requires the cooperation of several computers. These computers and their resources are distributed over the network. The resources needed for the execution of a task are allocated such that system load and the task delay incurred during its execution are optimized.<<ETX>>

Loadflow analysis on parallel computers

The speed performance of parallel computers in power system loadflow analysis is evaluated. Three commercial parallel computers, each of which has a fundamentally different architecture, are evaluated. A methodology for determining a suitable architecture for a given problem is developed. Major issues that cause performance degradation in parallel processing are discussed.<<ETX>>